3 Ways to Run CATIA Remotely: Which One Is Best?

A practical look at the three ways engineering teams give designers remote access to CATIA, and how each one holds up under real design and assembly work.   

Engineering teams rarely sit in one building anymore. Designers work from home, from a second site, or from a client location, and they still need full CATIA on the screen in front of them. For most business software, remote access is simple. CATIA is harder, because the experience depends on the hardware doing the work and on a licensing model that behaves in specific ways once you leave the office.  

CATIA runs on Windows and expects a certified professional graphics card and validated drivers, a high-performance CPU, and a large amount of memory. It also depends on a Dassault Systèmes License Server (DSLS) to hand out licenses, along with shared settings and PLM data that usually live on the office network. A remote setup that ignores any of these tends to disappoint.  

Newer releases raise the bar further. CATIA R2026x leans into cloud-connected roles, AI assistance, and Virtual Twin work, which means heavier graphics and more reliance on a well-specified environment. As the software grows more demanding, an underpowered remote connection falls further behind. 

When teams take CATIA out of the office, they usually pick one of three approaches: a VPN back to the office network, a remote desktop connection into an office workstation, or moving CATIA to the cloud. The cloud itself is not one thing, so we break it into its real routes further down. The sections below explain how each approach works, where each one struggles, and why a GPU-accelerated managed cloud desktop is generally the strongest fit for professional CATIA work. 

 If your specific blocker is Apple hardware rather than remote access, we cover that case separately in how to run CATIA on a Mac. 

The Three Ways Teams Run CATIA Remotely 

The table below summarizes the three approaches before we look at each one in detail. 

Running CATIA Over a VPN 

A VPN is often the first thing teams try because it is simple to set up. CATIA keeps running on the local machine, and the VPN connects it back to the office network to reach the DSLS license server and the design files. 

For licensing to work, that license server has to stay reachable. GoEngineer CATIA installation guidance recommends keeping the license server onsite and having remote users connect through the company VPN before launching CATIA.  If the VPN drops or the license server becomes unreachable, CATIA cannot check out or hold a license, so users may be interrupted mid-session or unable to launch until the connection is restored. 

The bigger limit is file transfer. CATIA assemblies and CATProduct structures are large, and every open and save moves that data across the tunnel. As the distance between the office and the remote user grows, each transfer takes longer. Files open slowly, saves lag, and without proper PLM control, two users can overwrite each other’s work. 

Shared settings add another wrinkle. When CATSettings or standards folders live on the office network and are not reliably reachable over the VPN, engineers see inconsistent behavior between home and office. 

A VPN can work for light part editing on a strong connection. It becomes difficult once large assemblies and surfacing are involved. 

Running CATIA Through Remote Desktop 

A remote desktop connection changes where CATIA runs. Instead of the local machine, CATIA runs on a workstation at the office, close to the license server and the design files. The remote user connects to that workstation, and only the screen image, mouse input, and keyboard input travel across the connection. 

That sounds like a clean fix, and for keeping files close it is. But CATIA introduces two problems that are specific to this method. 

The first is licensing. A CATIA nodelock license is tied to a physical machine and checks how you are connected. TECHNIA, a Dassault Systèmes partner, explains that a nodelock license cannot be acquired over a remote display, and the user has to be physically logged into the machine to start CATIA. A nodelock seat, in other words, does not launch over a standard remote desktop session. Teams that want to work this way need network licensing from a DSLS server instead, which adds setup and administration. 

The second is performance. Dassault partner Inceptra advises that Remote Desktop is not recommended for production CATIA work and is better suited to light parts, small assemblies, or review tasks. Rotating and panning large models over a standard remote session can lag, and some macros do not run remotely because of how shared resources are resolved across the connection. 

A 3D input device such as a SpaceMouse, which many CATIA users rely on all day, often behaves poorly over a basic remote connection. And the office workstation still has to exist for every user, stay powered on, and be maintained. If it fails, the user cannot work until someone is on site. 

Remote desktop solves the file problem by keeping a full workstation per person. That becomes costly and harder to manage as the team grows. 

Running CATIA in the Cloud 

Running CATIA in the cloud can mean three different things, and they are easy to mix up. The routes differ in what they replace, who operates them, and whether they keep your current CATIA, so it helps to separate them before deciding. 

Dassault 3DEXPERIENCE Cloud, the Official SaaS Route 

Dassault Systèmes sells its own cloud option. 3DEXPERIENCE CATIA on the cloud is a subscription service where CATIA is delivered as roles on Dassault’s platform, with data managed centrally rather than in local files. 

Two points determine whether it fits. First, this is the 3DEXPERIENCE generation, so for a team on CATIA V5 it is a platform move, not a way to run existing V5 work remotely. Dassault also offers CATIA V5 with cloud services, but there the V5 software still runs on the local machine, and only the data and collaboration layer sits in the cloud. 

Second, the heavy design apps are not fully browser-based yet. A growing set of roles runs in a browser, and CATIA R2026x added more web-compatible roles, but the full parametric design applications still install and run on a certified Windows machine, as Dassault’s hardware and software requirements show. So, the official SaaS route suits new projects, browser-based collaboration, and teams ready to move to the platform. It is a weaker fit for running your current CATIA V5 as is. 

Self-Managed Cloud on AWS or Azure 

The second route is to rent GPU instances from AWS or Azure and build the virtual desktop yourself. This keeps your existing CATIA and gives you full control, but the control comes with the work. 

Someone on your team sizes the GPU machines, keeps graphics drivers on the Dassault-validated list, tunes the streaming, patches Windows, runs the DSLS license server, and scales seats as projects change. Dassault updates its certified configurations regularly, so keeping the environment aligned is ongoing. For a team without spare cloud engineering time, this usually becomes heavier than expected. 

Managed Cloud VDI (Desktop as a Service) 

The third route keeps everything the self-managed option offers and removes the operational load. A desktop as a service provider runs the environment for you. Your existing CATIA, whether V5 or a 3DEXPERIENCE native app, runs on a GPU-powered Windows machine in a data center, and users connect from a laptop, a Mac, or a thin client. 

Only display data reaches the user screen, so the CAD data stays in the data center. The provider maintains the machine, keeps drivers aligned with Dassault-certified levels, tunes the streaming, runs the license server, scales seats, and supports the platform. Ace Cloud Hosting delivers this through Managed Citrix DaaS Solutions with 24×7 monitoring. Your team opens CATIA and works, with no office workstation to buy, power on, or replace at end of life. 

This is also the cloud route Dassault Systèmes validates directly. Dassault publishes certified configurations for virtual desktop infrastructure, and NVIDIA worked with Dassault to certify CATIA for VDI deployment, noting that running it this way removes the need to install CATIA on each local client and lowers IT support effort. 

The trade-off is the same for any cloud desktop: it depends on your internet connection, so you should weigh connection quality and latency. For most broadband and office links, this is a non-issue, but a weak or high-latency connection will be felt. 

How GPU Acceleration Supports CATIA Performance 

CATIA renders models using the graphics card rather than the main processor. The GPU handles rotating large assemblies, working with complex surfaces, and reviewing digital mockups. A hosted desktop meant for CATIA therefore cannot rely on a standard office-grade cloud machine. It needs a GPU-accelerated virtual desktop backed by an allocated NVIDIA vGPU profile. 

Ace Cloud Hosting delivers these desktops through Managed Citrix DaaS Solutions. The GPU resource is the allocated NVIDIA vGPU profile, and the delivery layer is Citrix HDX 3D Pro, which renders the 3D model on the server GPU and streams the resulting image to the user screen. Citrix is a certified delivery partner for the Dassault Systèmes 3DEXPERIENCE platform, so the protocol path is vendor-recognized. HDX 3D Pro was built for graphics-heavy applications and keeps viewport interaction responsive even over standard broadband, so a cloud-hosted desktop stays usable without an office-grade network. 

Support for 3D Input Devices 

A SpaceMouse is central to how many engineers pan, zoom, and rotate CATIA models. Many basic remote tools handle these devices poorly. Citrix HDX 3D Pro addresses this with USB redirection that prioritizes the 3D mouse channel, so a SpaceMouse works inside the hosted session much as it does on a local machine. 

Licensing and PLM Data in a Hosted Environment 

Licensing is often the hardest part of any remote CATIA deployment, as the remote desktop section showed. A cloud-hosted desktop avoids most of it by keeping the DSLS license server in the same environment as the desktops that run CATIA. They talk over a local, low-latency network instead of the public internet, so license checks stay stable. Depending on the terms of your Dassault agreement, you can usually keep your existing licenses. 

The same applies to data. Models, assemblies, and drawings stay in one central location rather than scattered across laptops. Teams using ENOVIA or another PLM system can keep the PLM platform and the CATIA desktops on the same backend, so check-in and check-out stay fast and version control stays organized. 

Matching the Right Approach to Your Team 

The table below matches each approach to a common situation. 

Data Security Considerations 

For engineering and manufacturing firms, design files are among the most valuable assets they hold. A cloud-hosted desktop is often more secure than a local machine. 

CATIA files never sit on the local device. They stay in the data center, and only an encrypted image of the screen reaches the user. If a laptop is lost or stolen, no design data goes with it. Administrators can also restrict local downloads, clipboard transfers, USB redirection, and printing to keep engineering data inside the managed environment. 

The data sits in data centers that maintain recognized security and compliance standards, with controls such as continuous monitoring, controlled physical access, and surveillance. Ace Cloud Hosting delivers the managed service on top of that certified data center infrastructure, so the compliance posture comes from the underlying facilities rather than from a single provider claim. 

Choosing the Right Way to Run CATIA Remotely 

A VPN is simple but struggles with large assemblies and depends on a stable link to the license server. Remote desktop keeps files close but runs into CATIA licensing behavior and weak graphics over a standard session, and it still needs a workstation for every user. 

The cloud is not a single choice either. Dassault’s own 3DEXPERIENCE cloud suits teams moving to the platform, self-managed GPU instances suit teams that want to build and run everything themselves, and managed cloud-hosted VDI gives each engineer a fast, fully specified environment running their existing CATIA, on the device they already use, with no infrastructure to build or maintain. 

The wider shift supports this. The desktop virtualization market is projected to grow from $33.9 billion in 2025 to $174.29 billion by 2034, as more teams move demanding workloads onto hosted desktops. For distributed engineering teams, heavy assemblies, or an approaching hardware refresh, CATIA Cloud Hosting on an allocated NVIDIA vGPU profile is generally the best match for how CATIA work is done today.